While federated learning is a promising approach for training deep learning models over distributed sensitive datasets, it presents new challenges for machine learning, especially when applied in the medical domain where multi-centric data heterogeneity is common. Building on previous domain adaptation works, this paper proposes a novel federated learning approach for deep learning architectures via the introduction of local-statistic batch normalization (BN) layers, resulting in collaboratively-trained, yet center-specific models. This strategy improves robustness to data heterogeneity while also reducing the potential for information leaks by not sharing the center-specific layer activation statistics. We benchmark the proposed method on the classification of tumorous histopathology image patches extracted from the Camelyon16 and Camelyon17 datasets. We show that our approach compares favorably to previous state-of-the-art methods, especially for transfer learning across datasets.
Semantic segmentation often requires a large set of images with pixel-level annotations. In the view of extremely expensive expert labeling, recent research has shown that the models trained on photo-realistic synthetic data (e.g., computer games) with computer-generated annotations can be adapted to real images. Despite this progress, without constraining the prediction on real images, the models will easily overfit on synthetic data due to severe domain mismatch. In this paper, we novelly exploit the intrinsic properties of semantic segmentation to alleviate such problem for model transfer. Specifically, we present a Regularizer of Prediction Transfer (RPT) that imposes the intrinsic properties as constraints to regularize model transfer in an unsupervised fashion. These constraints include patch-level, cluster-level and context-level semantic prediction consistencies at different levels of image formation. As the transfer is label-free and data-driven, the robustness of prediction is addressed by selectively involving a subset of image regions for model regularization. Extensive experiments are conducted to verify the proposal of RPT on the transfer of models trained on GTA5 and SYNTHIA (synthetic data) to Cityscapes dataset (urban street scenes). RPT shows consistent improvements when injecting the constraints on several neural networks for semantic segmentation. More remarkably, when integrating RPT into the adversarial-based segmentation framework, we report to-date the best results: mIoU of 53.2%/51.7% when transferring from GTA5/SYNTHIA to Cityscapes, respectively.
Conventional approaches to image de-fencing use multiple adjacent frames for segmentation of fences in the reference image and are limited to restoring images of static scenes only. In this paper, we propose a de-fencing algorithm for images of dynamic scenes using an occlusion-aware optical flow method. We divide the problem of image de-fencing into the tasks of automated fence segmentation from a single image, motion estimation under known occlusions and fusion of data from multiple frames of a captured video of the scene. Specifically, we use a pre-trained convolutional neural network to segment fence pixels from a single image. The knowledge of spatial locations of fences is used to subsequently estimate optical flow in the occluded frames of the video for the final data fusion step. We cast the fence removal problem in an optimization framework by modeling the formation of the degraded observations. The inverse problem is solved using fast iterative shrinkage thresholding algorithm (FISTA). Experimental results show the effectiveness of proposed algorithm.
Product descriptions in e-commerce platforms contain detailed and valuable information about retailers assortment. In particular, coding promotions within digital leaflets are of great interest in e-commerce as they capture the attention of consumers by showing regular promotions for different products. However, this information is embedded into images, making it difficult to extract and process for downstream tasks. In this paper, we present an end-to-end approach that classifies promotions within digital leaflets into their corresponding product categories using both visual and textual information. Our approach can be divided into three key components: 1) region detection, 2) text recognition and 3) text classification. In many cases, a single promotion refers to multiple product categories, so we introduce a multi-label objective in the classification head. We demonstrate the effectiveness of our approach for two separated tasks: 1) image-based detection of the descriptions for each individual promotion and 2) multi-label classification of the product categories using the text from the product descriptions. We train and evaluate our models using a private dataset composed of images from digital leaflets obtained by Nielsen. Results show that we consistently outperform the proposed baseline by a large margin in all the experiments.
We propose a training and evaluation approach for autoencoder Generative Adversarial Networks (GANs), specifically the Boundary Equilibrium Generative Adversarial Network (BEGAN), based on methods from the image quality assessment literature. Our approach explores a multidimensional evaluation criterion that utilizes three distance functions: an $l_1$ score, the Gradient Magnitude Similarity Mean (GMSM) score, and a chrominance score. We show that each of the different distance functions captures a slightly different set of properties in image space and, consequently, requires its own evaluation criterion to properly assess whether the relevant property has been adequately learned. We show that models using the new distance functions are able to produce better images than the original BEGAN model in predicted ways.
In optical coherence tomography (OCT) volumes of retina, the sequential acquisition of the individual slices makes this modality prone to motion artifacts, misalignments between adjacent slices being the most noticeable. Any distortion in OCT volumes can bias structural analysis and influence the outcome of longitudinal studies. On the other hand, presence of speckle noise that is characteristic of this imaging modality, leads to inaccuracies when traditional registration techniques are employed. Also, the lack of a well-defined ground truth makes supervised deep-learning techniques ill-posed to tackle the problem. In this paper, we tackle these issues by using deep reinforcement learning to correct inter-frame movements in an unsupervised manner. Specifically, we use dueling deep Q-network to train an artificial agent to find the optimal policy, i.e. a sequence of actions, that best improves the alignment by maximizing the sum of reward signals. Instead of relying on the ground-truth of transformation parameters to guide the rewarding system, for the first time, we use a combination of intensity based image similarity metrics. Further, to avoid the agent bias towards speckle noise, we ensure the agent can see retinal layers as part of the interacting environment. For quantitative evaluation, we simulate the eye movement artifacts by applying 2D rigid transformations on individual B-scans. The proposed model achieves an average of 0.985 and 0.914 for normalized mutual information and correlation coefficient, respectively. We also compare our model with elastix intensity based medical image registration approach, where significant improvement is achieved by our model for both noisy and denoised volumes.
Global warming is rapidly reducing glaciers and ice sheets across the world. Real time assessment of this reduction is required so as to monitor its global climatic impact. In this paper, we introduce a novel way of estimating the thickness of each internal ice layer using Snow Radar images and Fully Convolutional Networks. The estimated thickness can be analysed to understand snow accumulation each year. To understand the depth and structure of each internal ice layer, we carry out a set of image processing techniques and perform semantic segmentation on the radar images. After detecting each ice layer uniquely, we calculate its thickness and compare it with the available ground truth. Through this procedure we were able to estimate the ice layer thicknesses within a Mean Absolute Error of approximately 3.6 pixels. Such a Deep Learning based method can be used with ever-increasing datasets to make accurate assessments for cryospheric studies.
Current crowd counting algorithms are only concerned about the number of people in an image, which lacks low-level fine-grained information of the crowd. For many practical applications, the total number of people in an image is not as useful as the number of people in each sub-category. E.g., knowing the number of people waiting inline or browsing can help retail stores; knowing the number of people standing/sitting can help restaurants/cafeterias; knowing the number of violent/non-violent people can help police in crowd management. In this paper, we propose fine-grained crowd counting, which differentiates a crowd into categories based on the low-level behavior attributes of the individuals (e.g. standing/sitting or violent behavior) and then counts the number of people in each category. To enable research in this area, we construct a new dataset of four real-world fine-grained counting tasks: traveling direction on a sidewalk, standing or sitting, waiting in line or not, and exhibiting violent behavior or not. Since the appearance features of different crowd categories are similar, the challenge of fine-grained crowd counting is to effectively utilize contextual information to distinguish between categories. We propose a two branch architecture, consisting of a density map estimation branch and a semantic segmentation branch. We propose two refinement strategies for improving the predictions of the two branches. First, to encode contextual information, we propose feature propagation guided by the density map prediction, which eliminates the effect of background features during propagation. Second, we propose a complementary attention model to share information between the two branches. Experiment results confirm the effectiveness of our method.
Automatic image annotation is among the fundamental problems in computer vision and pattern recognition, and it is becoming increasingly important in order to develop algorithms that are able to search and browse large-scale image collections. In this paper, we propose a label propagation framework based on Kernel Canonical Correlation Analysis (KCCA), which builds a latent semantic space where correlation of visual and textual features are well preserved into a semantic embedding. The proposed approach is robust and can work either when the training set is well annotated by experts, as well as when it is noisy such as in the case of user-generated tags in social media. We report extensive results on four popular datasets. Our results show that our KCCA-based framework can be applied to several state-of-the-art label transfer methods to obtain significant improvements. Our approach works even with the noisy tags of social users, provided that appropriate denoising is performed. Experiments on a large scale setting show that our method can provide some benefits even when the semantic space is estimated on a subset of training images.
We present a conceptually simple but effective funnel activation for image recognition tasks, called Funnel activation (FReLU), that extends ReLU and PReLU to a 2D activation by adding a negligible overhead of spatial condition. The forms of ReLU and PReLU are y = max(x, 0) and y = max(x, px), respectively, while FReLU is in the form of y = max(x,T(x)), where T(x) is the 2D spatial condition. Moreover, the spatial condition achieves a pixel-wise modeling capacity in a simple way, capturing complicated visual layouts with regular convolutions. We conduct experiments on ImageNet, COCO detection, and semantic segmentation tasks, showing great improvements and robustness of FReLU in the visual recognition tasks.